Comments (8)
Yes, you can follow the same steps shown in this issue #6
from swiftwhisper.
Yes, you can follow the same steps shown in this issue #6
Thank you, I have however already tried. I should have been more clear.
It works great for continuous speaking audios.
But it works really bad when there are pauses in the speech, is there anything else than those parameterers I can work with?
from swiftwhisper.
Can you expand on this more? How does it work "really bad" for pauses in speech? What changes?
from swiftwhisper.
Can you expand on this more? How does it work "really bad" for pauses in speech? What changes?
If I make a recording when continually speaking it works great. If I make a recording where I say something like
"Okay here I'm speaking about" and then I take a pause of like 10 seconds and continue. Then in that 10 second gap I expect it to not show any word but it does come words but in a slower pace. And if I stop speaking and then let it run for some seconds in the end it will also add words in that space.
Continuous speaking good
Breaks are filled with words and then accelerates back inte correct place when continuous speaking are reestablished
from swiftwhisper.
I haven't tested it, but have you looked into using params.no_speech_thold
? Thats a value inside of whisper.cpp that is likely related to what you're looking for.
I don't think this is a bug inside of SwiftWhisper, so I'm going to keep this issue closed. Please re-open it if this a bug unique to this wrapper.
from swiftwhisper.
I haven't tested it, but have you looked into using
params.no_speech_thold
? Thats a value inside of whisper.cpp that is likely related to what you're looking for.I don't think this is a bug inside of SwiftWhisper, so I'm going to keep this issue closed. Please re-open it if this a bug unique to this wrapper.
Unfortunately it seems to make no difference.
Its like it won't add any pauses at all actually.
In this example I make many long pauses, but as you can see the next word always starts where the previous ends. No gaps at all? Is this another parameter that I have forgotten?
0 - 120 :
120 - 400 : So
400 - 800 : if
800 - 1420 : I
1420 - 1790 : make
1790 - 1990 : a
1990 - 3000 : video
3000 - 4980 : and
4980 - 5580 : if
5580 - 6570 : I
6570 - 7040 : take
7040 - 7160 : a
7160 - 7360 : long
7360 - 7620 : pause
7620 - 7960 : like
7960 - 10070 : that
from swiftwhisper.
@eppee I haven't tested, but just released version 1.0.3 which uses the latest version of whisper.cpp. Might be worth checking if this is still a problem.
from swiftwhisper.
@eppee I haven't tested, but just released version 1.0.3 which uses the latest version of whisper.cpp. Might be worth checking if this is still a problem.
Unfortunately there is no difference.
Been reading a lot and this seems to be a solution. But I don't know how to implement it in Swift...
from swiftwhisper.
Related Issues (20)
- memory free up issue HOT 1
- Confidence value for segment? HOT 3
- GPU vs CPU HOT 1
- Core ML Support HOT 2
- MLModelAsset: load failed with error HOT 2
- Are there plans to support more parameters in WhisperParams?
- Add support for word-level timestamps `--word_timestamps` HOT 3
- Is there a possibility to pause and resume transcription? HOT 2
- Running compiled C++ version of program is faster than using Library HOT 2
- Do you have more boilerplate code (examples) to show an example of some application where the speech recognition is working? HOT 6
- Missing required module 'whisper_cpp' when using the SwiftWhisper code. HOT 4
- Setting --max-len does not change the value of the results HOT 1
- watchOS support? HOT 7
- crash when initializing with an invalid model HOT 4
- More boilerplate besides the test functions. HOT 3
- thoughts on reading from microphone data? HOT 3
- Split segment to words HOT 1
- "loading Core ML model"Spend a lot of time
- large-v3 HOT 1
- Progress delegate is not working HOT 2
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from swiftwhisper.